Europe at a crossroads (Part 1 of 2)
Europe has lost the race to develop the world's most advanced AI models. But that does not mean we have lost the opportunity to build a competitive future with AI.
Artificial intelligence is already changing how we work, how companies operate and how knowledge is created and used. Its influence is spreading through education, healthcare, scientific research and the wider economy. It is also becoming an increasingly important factor in the global balance of power.
And AI is developing alongside advances in robotics, biotechnology, quantum computing and energy technology. Progress in one field can open up possibilities in another, accelerating changes that businesses, institutions and societies are already struggling to keep up with.
I see this as something bigger than an AI revolution. We are going through a technological shift that is changing some of the fundamental conditions under which people live, work and create value. We cannot yet know how far those changes will go, but the transformation is already under way.
A race Europe has lost
The United States is leading the development of the world's most advanced AI models, with China as its strongest competitor. Europe has fallen so far behind that I do not believe we will close the gap.
The investment figures tell an important part of the story. In 2025, private AI investment reached close to $286 billion in the United States, compared with just $21 billion in Europe.
Developing frontier models requires enormous amounts of capital, computing power, energy, specialised talent and infrastructure. The United States has built an investment and commercial ecosystem capable of supporting that work at a scale Europe has never matched. China has developed considerable capabilities of its own through research, industrial development and state-backed investment.
Europe has talented researchers and promising AI companies, but we are competing against an extraordinary concentration of resources that continues to grow. And the companies leading this development are hardly standing still.
We need to recognise what this means for our future. Europe will remain dependent on American, and to some extent Chinese, AI technology for the most demanding applications, including advanced scientific research and industrial development.
As AI becomes more deeply embedded in our economy, that dependence will matter. Decisions made by foreign technology companies and governments will influence which capabilities European businesses and institutions can access, under what conditions and at what cost.
I still believe we should continue developing our own AI technology. There are good reasons to invest in European models, research, infrastructure and technical expertise, even if our models never become the most advanced in the world.
Many applications simply do not require frontier-level capabilities. Smaller or more specialised models may be perfectly capable of handling internal documentation, information retrieval, routine analysis and other well-defined business tasks.
Where European technology can meet those needs, it gives businesses more choice and greater control over their data and infrastructure. It also allows us to retain expertise and develop technology around our own requirements.
Complete independence across the entire AI technology stack is unrealistic. We will continue relying on foreign technology, particularly for the most demanding applications. But there is a considerable difference between depending on foreign models for selected tasks and allowing that dependence to extend across almost everything.
For me, technological sovereignty is about retaining as much meaningful control as possible. We need to know which capabilities we depend on, develop credible alternatives where we can and avoid surrendering control unnecessarily.
Where Europe should put its energy
Much of the attention surrounding AI goes to the companies developing the models. Their progress is extraordinary, and the technology they produce will influence almost every industry.
But there is another part of this competition where I believe Europe should concentrate far more of its efforts: helping businesses become exceptionally good at using AI.
We already know how to build sophisticated industrial companies. Across Europe, businesses have spent decades developing specialist knowledge of their products, processes and markets. Manufacturing, energy, engineering and healthcare are just some of the industries where we have accumulated considerable expertise.
I am particularly interested in what happens when that knowledge becomes easier to use throughout a company.
Imagine a manufacturer where employees can draw on previous technical experience without having to know exactly who holds that information or where it is stored. Where recurring problems in production lead to improvements that are retained and made available to others. Where routine administrative work can be handled through connected systems, allowing experienced employees to spend more time applying their expertise.
These possibilities matter because much of Europe's economic strength already exists inside established businesses. We do not have to invent entirely new industries to benefit from AI. There is enormous potential in helping the companies we already have operate differently.
A company can make individual employees more productive while leaving its underlying operations largely unchanged. Knowledge remains scattered, processes continue as before, and useful methods developed by one person may never become available to anyone else.
The larger gains come when companies reconsider how work is organised and make AI part of their core operations.
Finding the right balance between regulation and innovation
One of my biggest concerns is that Europe could make this transformation unnecessarily difficult through its own approach to regulation.
We need rules governing the development and use of artificial intelligence. As systems become more capable and gain access to sensitive information, business processes and consequential decisions, clear responsibilities and appropriate safeguards are essential.
There are good reasons to protect privacy, safety and individual rights. Companies should be accountable for how they use AI, particularly when it affects people's employment, access to services or other important aspects of their lives.
My concern is what happens when those requirements become so complicated that businesses struggle to understand how to comply with them.
Europe already has a reputation for extensive bureaucracy. With AI developing as quickly as it is, the practical consequences of unclear or excessively demanding regulation deserve serious attention.
Consider a small logistics company that wants to introduce AI into delivery planning and routine customer enquiries. It may have identified a useful application and be willing to invest in it, but it probably has no dedicated legal department and only limited technical resources.
Its management needs to understand which rules apply, what it can do with its data and what safeguards are required. If those answers are difficult to obtain, and the consequences of an unintentional mistake are unclear, postponing the project may seem like the safest option.
That is a serious problem for an economy that depends so heavily on smaller and medium-sized businesses.
A system helping employees organise internal documentation presents different risks from one making consequential decisions about employment or access to credit. The rules should reflect those differences, and companies should be able to understand their obligations without needing extensive legal assistance for every new application.
Europe has recognised some of these difficulties. The Apply AI Strategy aims to accelerate adoption, while the AI Omnibus reforms that entered into force in July 2026 introduced simplifications, extended certain implementation deadlines and expanded opportunities for testing.
I welcome these developments, although their value will ultimately be determined by what businesses experience in practice. A company that still cannot establish what it is allowed to do is unlikely to find much reassurance in another European strategy document.
I want Europe to maintain high standards for the responsible use of AI. I also want companies to feel confident enough to experiment, invest and develop new capabilities.
We are already behind in the development of advanced AI. We cannot afford to create unnecessary obstacles for the businesses trying to put that technology to work.
Giving businesses a reason to adopt AI
After the Second World War, Europe faced serious challenges around agricultural productivity and food security. European countries recognised the importance of modernising agriculture and increasing production.
The Common Agricultural Policy, introduced in 1962, established shared incentives for investment and greater output. Its support mechanisms eventually contributed to surpluses, market distortions and substantial costs that required reform.
But what interests me is the ambition behind the policy. Europe recognised that developing a particular economic capability was strategically important and was prepared to act collectively to strengthen it.
I think artificial intelligence deserves a similar level of attention today.
Many smaller businesses understand that AI is becoming important. That does not mean they know where it could create value, how to introduce it into their operations or who should maintain the systems afterwards.
A company might see opportunities to improve production planning, reduce administrative work or develop better customer services. Turning those ideas into functioning systems takes time, expertise and investment. And for a business with limited resources, the uncertainty can be enough to prevent it from getting started.
Europe could do much more to lower these barriers. Practical technical support, employee training, access to expertise and financial incentives could help businesses make investments they would otherwise struggle to justify.
I would like to see that support directed towards companies making substantial improvements to how they operate, rather than simply rewarding the purchase of more AI software.
Consider the difference between subsidising another set of AI licences and helping a manufacturer redesign an expensive production process. The first gives employees access to a tool. The second could leave the business with a better way of working that continues to deliver value.
We should also take the costs of implementation seriously. New systems require maintenance, employees need training, and established processes often have to change before improvements become visible. For many smaller businesses, access to practical assistance will matter just as much as the technology itself.
The Common Agricultural Policy showed that European incentives can change investment decisions across an entire sector, although it also demonstrated the consequences of rewarding output without sufficient regard for its economic value.
A different European industrial ambition
Europe's Industry 5.0 perspective offers a useful starting point for thinking about the kind of businesses we should be building.
Industry 4.0 concentrated heavily on digitalisation, automation and more efficient production. Industry 5.0 builds on those foundations while giving greater attention to people, resilience and sustainability.
What interests me is how those ideas might change everyday decisions inside a business.
Take a manufacturer introducing AI into its production processes. It could focus narrowly on reducing labour costs or increasing output. Those may be reasonable objectives, but they leave a much larger question unanswered.
Could the company reduce waste by identifying production problems earlier? Could experienced technicians make their knowledge available to colleagues without having to explain the same things repeatedly? Could employees spend less time compiling reports and more time solving problems that require their judgement? And could the business respond to changing demand without continuously adding more people and administrative layers?
AI creates opportunities to rethink the relationship between the resources a business uses and the value it produces. A company may be able to increase its capacity without increasing costs at the same rate. It might become less dependent on particular employees for essential information, or make better use of energy and materials.
This has implications for competitiveness, but also for the people working in these businesses. Some responsibilities will change substantially as technology becomes more capable. Employees will need opportunities to develop new skills, and companies will have to decide where human expertise, creativity and judgement make the greatest contribution.
I do not see Industry 5.0 as a reason to slow down technological development. I see it as a challenge to make better use of what technology allows us to do.
A business that produces more with fewer resources, retains the knowledge it develops and gives its employees better ways to perform their work is worth building. That is a much more tangible ambition than simply announcing that a company has become AI-driven.
The position Europe chooses
Other economies are already investing heavily in advanced AI, and businesses around the world are experimenting with new ways to organise work and develop products. As those approaches mature, the pressure on European companies will increase.
We have decisions to make about research, infrastructure, regulation and the incentives we create for businesses. These decisions matter because they will influence whether European companies can keep pace with the changes already taking place elsewhere.
What concerns me most is the possibility that we become a continent that purchases increasingly advanced AI technology from abroad without becoming particularly good at using it.
We have companies with decades of industrial experience, substantial technical knowledge and people who understand some of the world's most demanding industries. I would rather see us use those strengths to build better businesses than spend the coming years trying to recreate the American AI industry on a smaller scale.
What we do with the technology being developed will determine how much further behind we fall, or whether we can establish a different kind of strength.